Vapor phase Beckmann rearrangement of cyclohexanone oxime on Hβ-zeolites treated by ammonia

Vapor phase Beckmann rearrangement of cyclohexanone oxime on Hβ-zeolites treated by ammonia

Available online at www.sciencedirect.com Microporous and Mesoporous Materials 107 (2008) 247–251 www.elsevier.com/locate/micromeso Vapor phase Beck...

226KB Sizes 1 Downloads 30 Views

Available online at www.sciencedirect.com

Microporous and Mesoporous Materials 107 (2008) 247–251 www.elsevier.com/locate/micromeso

Vapor phase Beckmann rearrangement of cyclohexanone oxime on Hb-zeolites treated by ammonia Yongjie Zhang, Yaquan Wang *, Yifeng Bu Key Laboratory for Green Chemical Technology, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, PR China Received 13 January 2007; received in revised form 8 March 2007; accepted 9 March 2007 Available online 24 March 2007

Abstract Hb was synthesized by a hydrothermal method followed by ion exchange and treated with aqueous solutions of ammonia. The obtained catalysts were characterized by BET surface area measurement, XRD, XPS, TEM, IR, pyridine-TPD and studied in the vapor phase Beckmann rearrangement of cyclohexanone oxime. The results show that in comparison with pure Hb, the ammonia treatments led to higher BET area, lower SiO2/Al2O3 ratio, more total acid sites, more weak acid sites, especially more weak Brønsted acid sites, and slow deactivation in the Beckmann rearrangement reaction. The results indicate that for the catalysts having the structure of b-zeolite, the suitable acid sites for the Beckmann rearrangement might be the weak Brønsted acid sites.  2007 Elsevier Inc. All rights reserved. Keywords: Hb-zeolite; Ammonia; Cyclohexanone oxime; Beckmann rearrangement; e-Caprolactam

1. Introduction The Beckmann rearrangement of cyclohexanone oxime is an important step in the manufacture of e-caprolactam, the monomer of nylon-6. Currently, the widely used technologies for the production of e-caprolactam produce large amounts of byproducts. The new process for the production of e-caprolactam is direct ammoximation of cyclohexanone with NH3 and H2O2 catalyzed by TS-1 [1–3], followed by vapor phase Beckmann rearrangement of the formed cyclohexanone oxime catalyzed by solid acid catalysts [4]. For the Beckmann rearrangement, many solid acid catalysts such as alumina [5], zeolites including Y [6], Silicalite-1 (S-1), those with MFI-structure [7–10] and b [11,12] have been studied. However, all the catalysts deactivate fast and the selectivity for e-caprolactam is low and no unanimous understanding about the reaction mechanism has been achieved. *

Corresponding author. Tel./fax: +86 22 27406335. E-mail address: [email protected] (Y. Wang).

1387-1811/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2007.03.015

b-Zeolite, first synthesized in 1967 [13], has a threedimensional interconnected channel system with 12-membered elliptical openings having mean diameters of 0.67 nm. The aluminum atoms in b-zeolite can produce Brønsted as well as Lewis acidity. The Brønsted acid sites are present both on the internal and external surface while Lewis sites are present on the internal surface [14]. b-Zeolite has high thermal and hydrothermal stability and novel catalytic properties. Much attention has been paid to bzeolite because of its unique structure, acidity and activity as solid acid catalysts. It was reported that Hb was a relatively highly active, selective and stable catalyst for the ecaprolactam formation from cyclohexanone oxime [11,12]. It was also reported that treating zeolites in alkaline solution could enlarge the aperture of zeolites or produce roughness on the outer surface and that the catalytic activities of the reaction between benzene and propylene, or between toluene and propylene, and the selective oxidation of cyclohexane, were greatly improved by such treatments [15,16]. In the present work, Hb was synthesized by a hydrothermal method followed by ion exchange and treated with

248

Y. Zhang et al. / Microporous and Mesoporous Materials 107 (2008) 247–251

aqueous solutions of ammonia. The obtained catalysts were characterized and studied in the vapor phase Beckmann rearrangement of cyclohexanone oxime. The aims of this work are to find out if this kind of treatments has effects on the reaction and to obtain some useful information about the active sites for the Beckmann rearrangement. 2. Experimental 2.1. Catalyst preparation b-Zeolite was synthesized under hydrothermal conditions and Hb-zeolite was obtained by ion exchange with ammonium chloride [17]. Hb was treated with ammonia solutions. In a typical run, 4.14 g H2O and 0.56 g ammonia solution (25 wt.%) was put into a PTFE-lined autoclave, then 1.00 g Hb-zeolite was added. The treatment was conducted at room temperature for 24 h under static conditions. Then the product was filtered, washed till the filtrate was neutral. The yielded material was dried at 100 C for 4 h and then calcined at 400 C for 5 h in air. Hb-NH3-0.56 g, Hb-NH3-3.05 g and Hb-NH3-6.09 g represent the catalysts with 1.00 g Hb being treated at room temperature with 0.56 g, 3.05 g and 6.09 g ammonia solution, respectively. 2.2. Catalyst characterization BET surface area was measured by N2 adsorption at 77 K with a QUANTA CHAROME CHEMBET-3000 apparatus. X-ray diffraction (XRD) was performed on a RIGAKUD/max-2500 X-ray instrument using Cu Ka radiation with a scan speed of 2 min 1. X-ray photoelectron spectroscopy (XPS) analysis was carried out on a PHI 1600 ESCA system by using Mg Ka radiation. Measurements were done at 29.35 eV pass energy and charging was corrected using the C 1s signal at 284.6 eV. Transmission electron microscopy (TEM) was carried out on a JEOL 100CXII apparatus. The FT-IR spectra of adsorbed pyridine were recorded on a Nicolet Magna 560 spectrometer and the acid strength distribution of the catalysts were determined by pyridine-TPD using TGA-50 (SHIMADZU) according to [17]. The amounts of coke over the deactivated catalysts were also measured using the above TGA instrument according to [17].

HPLC Pump at the rate of 6 cm3 h 1. 20 cm3 min 1 of Ar was introduced using a mass flow controller. The reaction products were collected in an ice–water trap and analyzed using an Agilent 4890GC with packed column (PEG-20M, 2 m) and a flame ionization detector (FID). The conversion of cyclohexanone oxime and the selectivity to e-caprolactam were calculated as follows: conversion (%) = (moles of cyclohexanone oxime in the feed-moles of unreacted cyclohexanone oxime in the product)/moles of cyclohexanone oxime in the feed · 100%; selectivity (%) = (moles of oxime in the product/moles of the product) · 100%. 3. Results and discussion The XRD patterns shown in Fig. 1 indicate that no dramatic breakage to the structure of zeolite b was caused by the treatment, although the basicity of the ammonia solutions are very high. The results of XPS analysis and BET areas are listed in Table 1. The results show that the ratio of SiO2/Al2O3 decreased to some extents after treatments, which might be caused by the removal of Si due to the effect of ammonia. The results also show that the BET areas of the ammonia treated Hb catalysts were higher than that of Hb. This may be attributed to the formation of cavities due to the removal of Si. Fig. 2 shows the TEM micrographs of pure Hb and the ammonia treated Hb catalysts. It is seen that Hb takes on regular figuration while the ammonia treated Hb catalysts have cavities and roughness. The results further indicate that some of the framework Si was removed by the treatment. The pyridine-TPD profiles on Hb and the ammonia treated Hb catalysts are given in Fig. 3. The amounts of strong and weak acid sites were calculated according to [17] and are listed in Table 1. The results show that the ammonia treatments markedly increased the weak acid sites, while the number of the strong acid sites decreased

d

2.3. Beckmann rearrangement reactions

c

The catalytic vapor phase Beckmann rearrangement reactions were carried out in a fixed-bed flow reactor. Five hundred milligram of catalyst was fixed with quarts wool in a quarts reactor (8 mm i.d.). The reactor was placed inside a temperature-controllable vertical furnace. The thermocouple tip in the quartz well was centered at the middle of the catalyst. After treated in a flow of Ar at 350 C for 1 h, 10 wt.% of cyclohexanone oxime in methanol was fed into the reactor after vaporization by a Series IV Digital

b a 5

10

15

20

25

30

35

ο

2θ ( ) Fig. 1. XRD patterns of Hb and the ammonia treated Hb catalysts. a: Hb catalyst, b: Hb-NH3-0.56 g catalyst, c: Hb-NH3-3.05 g catalyst, d: HbNH3-6.09 g catalyst.

Y. Zhang et al. / Microporous and Mesoporous Materials 107 (2008) 247–251

249

Table 1 Properties of Hb and the ammonia treated Hb catalysts before and after reaction Catalysts

BET area (m2 g 1)

SiO2/Al2O3 (mol ratio)

Total acid sites (lmol g 1)

Weak acid sites (lmol g 1)

Strong acid sites (lmol g 1)

Bronsted acid sites (lmol g 1)

Lewis acid sites (lmol g 1)

Coke deposited after reaction (%)

Hb Hb-NH3-0.56 g Hb-NH3-3.05 g Hb-NH3-6.09 g

438 476 484 504

30.35 24.32 24.06 23.16

761 1359 1503 1372

220 838 1134 875

541 521 369 497

462 740 735 731

299 619 768 641

8.61 9.84 10.74 9.59

0.4 mg)

Fig. 2. TEM micrographs of Hb and the ammonia treated Hb catalysts. A: Hb catalyst, B: Hb-NH3-0.56 g catalyst, C: Hb-NH3-3.05 g catalyst, D: Hb-NH3-6.09 g catalyst.

Weight (

a

The IR spectra of pyridine adsorption on Hb and the ammonia treated Hb catalysts are shown in Fig. 4. It is seen that both Brønsted and Lewis acids are present on Hb and the ammonia treated Hb catalysts. According to [17], the acidities of Brønsted and Lewis acids were calculated and the results are also listed in Table 1. The results demonstrate that the ammonia treatments resulted in the increase of both Brønsted and Lewis acid sites. All the catalysts treated by ammonia have similar amount of Brønsted acid sites. Increasing the amount of ammonia from 0.56 g to 3.05 g, the Lewis acid sites increased from 619 lmol g 1 to 768 lmol g 1. But further increasing the amount of ammonia from 3.05 g to 6.09 g did not cause further increase of the Lewis acid sites. Fig. 5 shows the changes of the conversion and selectivity of the Beckmann rearrangement of cyclohexanone oxime with time on steam over Hb-zeolite and the ammonia treated Hb catalysts. In the vapor phase Beckmann rearrangement of cyclohexanone oxime, all of the solid acid catalysts reported deactivated to some extents [18]. Fig. 5 shows that all the catalysts exhibited similar initial activity, but with ammonia treatment, the catalysts deactivated more slowly than the original Hb, and that the increase of ammonia amounts in the treatments increased the stability of the catalysts. The conversion of cyclohexanone oxime at 8 h on Hb-NH3-6.09 g was still 96.39%, while that on the original Hb dropped to 56.86%. The selectivities to caprolactam on the ammonia treated Hb catalysts were a bit lower than on Hb.

b c d

200

400

600

800

Fig. 3. Pyridine-TPD profiles on Hb and the ammonia treated Hb catalysts. a: Hb catalyst, b: Hb-NH3-0.56 g catalyst, c: Hb-NH3-3.05 g catalyst, d: Hb-NH3-6.09 g catalyst.

to some extent. Hb-NH3-3.05 g has the most weak acid sites and least strong acid sites, 1134 lmol g 1 and 369 lmol g 1, respectively. But further increasing the amount of the ammonia used in the treatment did not cause further change of the acidities. Hb-NH3-6.09 g has weak acid sites and strong acid sites, 875 lmol g 1 and 497 lmol g 1, respectively, which are almost equal to those of Hb-NH3-0.56 g catalyst.

Absorbance

Temperature (oC)

a b c d

1590

1560

1530

1500

1470

1440

1410

Wavenumbers (cm-1) Fig. 4. IR spectra of pyridine adsorption on Hb and the ammonia treated Hb catalysts. a: Hb, b: Hb-NH3-0.56 g, c: Hb-NH3-3.05 g, d: Hb-NH36.09 g.

250

Y. Zhang et al. / Microporous and Mesoporous Materials 107 (2008) 247–251

amounts of coke on the ammonia treated Hb catalysts were more than that on Hb. The results indicate that the more acid sites the catalysts have, the more coke might be formed after reaction. From the above discussion along with the report in the literature [19,20], we conclude that the weak Brønsted acid sites favor the Beckmann rearrangement. The reaction might proceed as follows: The weak Brønsted acid site first attacks the nitrogen atom of the oxime and then transfers it to the oxygen atom. The transfer of the alkyl group and the elimination of the hydroxyl group of the oxime as water, interacting with hydrogen of weak Brønsted acid site, simultaneously occur [21].

100

Conversion (%)

80

60

40

20

100

4. Conclusions

Selectivity (%)

80

The results show that the ammonia treatments of Hbzeolite led to higher BET area, lower SiO2/Al2O3 ratios, more total acid sites, more weak acid sites, especially more weak Brønsted acid sites, and increased the stability of the zeolite in the Beckmann rearrangement reaction. The results indicate that for the catalysts having the structure of b-zeolite, the suitable acid sites for the Beckmann rearrangement might be the weak Brønsted acid sites and the deactivation of the catalysts might be caused by the coke deposition.

60

40

Hβ Hβ-NH3-0.56 g Hβ-NH3-3.05 g

20

Hβ-NH3-6.09 g 0

0

2

4

6

8

10

Time (h) Fig. 5. Effects of ammonia concentrations in the treatment of Hb on the Beckmann rearrangement of cyclohexanone oxime.

Acknowledgment The work is supported by a State Key Fundamental Research Project of China (Grant No. G2000048005).

Weight (

0.5 mg)

References

d c b a 0

200

400

600

800

Temperature (oC) Fig. 6. TGA curves in measuring the amounts of coke on catalysts after Beckmann rearrangement reaction. a: Hb, b: Hb-NH3-0.56 g, c: Hb-NH33.05 g, d: Hb-NH3-6.09 g.

After 8 h on stream, the catalysts were collected and the deposited carbon was measured using TGA and the TGA curves are shown in Fig. 6. The amounts of coke were calculated and are also given in Table 1, which show that the

[1] G. Petrini, G. Leofanti, M.A. Mantegazza, F. Pognataro, in: P.T. Anastas, T.C. Williamson (Eds.), Caprolactam via Ammoximation, in Green Chemistry: Designing Chemistry for the Environment, American Chemical Society, Washington, DC, 1996, p. 33. [2] C.T. Wu, Y.Q. Wang, Z.T. Mi, L. Xue, W. Wu, E.Z. Min, S. Han, F. He, S.B. Fu, React. Kinet. Catal. Lett. 77 (2002) 73. [3] T.F. Liu, X.K. Meng, Y.Q. Wang, X.H. Liang, Z.T. Mi, X.J. Qi, S.Y. Li, W. Wu, E.Z. Min, S.B. Fu, Ind. Eng. Chem. Res. 43 (2004) 166. [4] K. Tanabe, W.F. Ho¨lderich, Appl. Catal. A: General 181 (1999) 399. [5] D.C. England, US 2634269, 1953. [6] P.S. Landis, P.B. Venuto, J. Catal. 6 (1966) 245. [7] H. Sato, K. Hirose, N. Ishii, Y. Umada, US 4709024, 1987. [8] J. Ro¨seler, G. Heitmann, W.F. Ho¨lderich, Appl. Catal. A: General 144 (1996) 319. [9] H. Ichihashi, H. Sato, Appl. Catal. A: General 221 (2001) 359. [10] W.F. Ho¨lderich, G. Dahlhoff, H. Ichihashi, US 6531595, 2002. [11] L.X. Dai, R. Hayasaka, Y. Iwaki, K.A. Koyano, T. Tatsumi, Chem. Commun. (1996) 1071. [12] G.P. Heitmann, G. Dahlhoff, W.F. Ho¨lderich, Appl. Catal. A: General 185 (1999) 99. [13] R.L. Wadlinger, G.T. Kerr, E.J. Rosinski, US 3308069, 1967. [14] J.C. Jansen, E.J. Creyghton, S.L. Njo, H. van Koningsveld, H. van Bekkum, Catal. Today 38 (1997) 205. [15] J. Liu, H.C. Guo, X.S. Wang, CN 1142986, 1997. [16] L. Zhang, A.Z. Ma, Z.M. Jin, CN 1358570, 2002. [17] Y.J. Zhang, Y.Q. Wang, Y.F. Bu, Z.T. Mi, W. Wu, E.Z. Min, S. Han, S.B. Fu, Catal. Commun. 6 (2005) 53.

Y. Zhang et al. / Microporous and Mesoporous Materials 107 (2008) 247–251 [18] T. Takahashi, M.N.A. Nasution, T. Kai, Appl. Catal. A: General 210 (2001) 339. [19] D. Shouro, Y. Moriya, T. Nakajima, S. Mishima, Appl. Catal. A: General 198 (2000) 275.

251

[20] K. Chaudhari, R. Bal, A.J. Chandwadkar, S. Sivasanker, J. Mol. Catal. A 177 (2002) 247. [21] G. Dahlhoff, J.P.M. Niederer, W.F. Ho¨lderich, Catal. Rev.-Sci. Eng. 43 (2001) 381.